For example, you might expect that a column indicating the country of origin for a product should not be null when the product is an import. If it’s fine for the country of origin column to be null for products produced locally, then applying a nullness check for country of origin on every row in the Data Asset could result in many false negatives.
Historically, the only way to express this level of nuance for your tests in GX Cloud was through custom SQL Expectations, requiring technical skill and maintenance to validate your data.
The latest improvements to Row conditions allow a user to combine multiple logic statements in a test with no SQL required. Each condition can be connected using intuitive AND/OR relationships when you set up an Expectation.
Condition Blocks: group related conditions using AND.
Multiple Blocks: connect blocks using OR.
New Set Operators: is any of and is none of for simpler list-based logic.
Let’s say we need to only apply a test within the “books,” “magazines,” and “cards” product categories where more than 100 of that category of products was purchased. The Row conditions below get us what we need.
What about monitoring how a new product launch is going where purchases OR returns happen after the launch date? It’s easier than ever to design tests that reflect your business’s needs.
Why this matters
Accessible to everyone: No need to know SQL to build powerful validations.
Scalable: Define and maintain rules in a structured, simple, and visual format.
Readable: Clear logic and description limits keep expectations easy to scan and audit.
Backward-compatible: Your existing expectations will work through API v1.8.0+, while upgrading allows you to use the new Condition object.
A smarter, simpler way to test your data
Providing multiple row conditions gives every user—engineer or analyst—the power to define logic that mirrors real business rules.
It’s the next step toward making GX Cloud the most user-friendly data-quality platform on the market.
Want to try it out? Start a trial of GX Cloud.
